Approximately 40,000 Americans are diagnosed with brain tumors each year, with15-35% being glioblastoma multiforme (GBM);the most aggressive primary brain tumor that has defied all existing treatment modalities. Treating brain tumors begins with surgical resection then follows with radiation or chemotherapy. Surgery faces the risks of removing surrounding tissues that may carry vital brain functions, while both radiation and chemotherapy can also harm normal tissues along the treatment pathway. Chemotherapy has been offering very limited applications, due to the palliative response and lack of targeting and selectivity of the drugs. Proposed herein is a novel drug delivery system (DDS) that will utilize MION (magnetic iron oxide nanoparticles) as the carrier to achieve synchronized MRI and drug therapy of brain tumors. It contains all desirable features within a single DDS including: [1] MRI, [2] magnetic targeting, [3] prodrug, and [4] cellular drug uptake, in overriding obstacles in brain drug delivery and achieving MRI-visualized, highly effective tumor therapy with least drug- induced toxic effects. In principle, macromolecular drug (e.g. ATF5-siRNA) with unmatched glioma specificity and potency will be linked to the non-toxic cell-penetrating LMWP via cytosol-degradable S-S bond, whereas MION carrying superparamagnetic behavior and superior magnetophoretic mobility will be coated with a bio- compatible heparin-dextran polymer. The LMWP-modified drug (LMWP-Drug) and heparin-coated MION (Hep- MION) will automatically group into a complex via electrostatic binding between the cationic LMWP and anionic heparin. After assembly, LMWP-Drug/Hep-MION shall display a unique prodrug feature during tumor targeting, due to inhibition of LMWP's trans-cell activity by heparin binding. To prevail over first-pass organ clearance thus maximizing MION accumulation at the tumor, the complexes will be injected via intra-arterial route. Optimized magnetic field topography will then follow to abort possible embolism of arterial vasculature and maximize tumor targeting selectivity. After tumor localization of MION via passive EPR- and active magnetic-targeting is verified by MRI, nasal administration of protamine, a clinical heparin antidote that binds heparin stronger than LMWP, will be followed to trigger release of LMWP-Drug from Hep-MION. Once inside tumor cells by LMWP-mediated internalization, the drug will be detached from LMWP by degradation of the S-S bond via elevated cytosolic reductase activity, initiating tumor apoptosis. Since large drugs are cell-impermeable, the cytosol-delivered drugs will not be affected by MDR. Preliminary findings were extremely promising, as they demonstrated by far the first true success of delivering a significant amount of the 465-KDa b-galactosidase selectively into the brain tumor but not ipsilateral or contralateral normal brain regions. In this new R01 application, we plan to confirm the utility of this DDS in vivo using well-established rat glioma models.